This proposal is the second phase of a prospective longitudinal study of alcohol use and other problem behaviors in a two-generational family study. The first phase of the study included a four-wave panel design of over 1,000 adolescents and their primary caregivers, with measurement occasions spaced at six-month intervals. The primary focus of the first phase was to use a lifespan developmental perspective to guide hypotheses about multiple and interacting vulnerability, or risk, factors that contribute to both externalizing (e.g., alcohol and substance use, delinquent activity) and internalizing (e.g., depressive symptoms, suicidal behaviors) domains, and a range of vulnerability factors were measured (e.g., family history of alcoholism, childhood behavior problems, temperament, family/friend support). The current proposal seeks to extend this multivariate developmental risk paradigm with a focus on role transitions in young adulthood (e.g., single-married, student-employed). Young adulthood has often been characterized as a period of instability of drinking behavior, with salient transitional events (e.g., getting married) impacting on drinking and other problem behaviors. Similar to the first phase of this project, multiple venerability factors and internalizing and externalizing problems will be measured to facilitate hypothesis-testing about alternative developmental pathways to problem behaviors in young adulthood for example, for men and women, for cumulative problem drinkers from adolescence onward and later onset problem drinkers). An expanded feature of this phase of the project is to focus on primary caregivers (primarily women) within the same multivariate developmental vulnerability factor/problem behavior model as the adolescents. This will facilitate the longitudinal study of a relatively large nonclinical sample of women, and enable the evaluation of ongoing intergenerational and bidirectional influences between primary caregivers and their adult children. Direct interviews will be used for the first wave of measurement in phase two, and a mail survey will be used one year later for the second wave. Data analytic procedures used to study the longitudinal relations will include structural equation modeling, log- linear and latent class modeling, and survival analysis.